A thoroughly interesting and enjoyable romp. Makes me wonder what happens if you ask LLM's how they know I'm alive - what questions would I be asked? I guess someone must have tried this?
yeah, that's actually quite a common technique that people use to try and wrestle with the beast - on the grounds that they are best placed to know about themselves...
The problem is it is hard to reason about the corpus that the LLM is drawing on - because maybe I have written something about you that you don't know about, or there is a commercially amplified version of you kicking about based on your data being processed and augmented in some way
Thanks - that’s a really interesting piece and a good read
Thanks Jackie - happy to answer any other questions/worries/concerns about it.
I am very much a lay user - like I managed an ML/AI team but I am not a node twiddler or matrix weighter myself...
A thoroughly interesting and enjoyable romp. Makes me wonder what happens if you ask LLM's how they know I'm alive - what questions would I be asked? I guess someone must have tried this?
yeah, that's actually quite a common technique that people use to try and wrestle with the beast - on the grounds that they are best placed to know about themselves...
The problem is it is hard to reason about the corpus that the LLM is drawing on - because maybe I have written something about you that you don't know about, or there is a commercially amplified version of you kicking about based on your data being processed and augmented in some way